Catalysis Science & Technology

0 downloads 0 Views 2MB Size Report
Apr 27, 2016 - The versatility and ease of application make this catalytic process concept .... The manufacturing of biodiesel by heterogeneous catalysis has .... method for designing a laboratory PFR reactor by employing similarity principles.
View Article Online

Catalysis Science & Technology

View Journal

Accepted Manuscript

This article can be cited before page numbers have been issued, to do this please use: A. C. Dimian and G. Rothenberg, Catal. Sci. Technol., 2016, DOI: 10.1039/C6CY00426A.

This is an Accepted Manuscript, which has been through the Royal Society of Chemistry peer review process and has been accepted for publication. Accepted Manuscripts are published online shortly after acceptance, before technical editing, formatting and proof reading. Using this free service, authors can make their results available to the community, in citable form, before we publish the edited article. We will replace this Accepted Manuscript with the edited and formatted Advance Article as soon as it is available. You can find more information about Accepted Manuscripts in the Information for Authors. Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal’s standard Terms & Conditions and the Ethical guidelines still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains.

www.rsc.org/catalysis

Page 1 of 20

Catalysis Science & Technology View Article Online

DOI: 10.1039/C6CY00426A

An effective modular process for biodiesel manufacturing using heterogeneous catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Alexandre C. Dimian*, Gadi Rothenberg Van’t Hoff Institute for Molecular Sciences, University of Amsterdam P.O. Box 94157, 1090GD Amsterdam, The Netherlands [email protected] (corresponding author), [email protected]

15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

We present an innovative reaction set-up and process for biodiesel manufacturing by heterogeneous catalysis. This process has two key advantages over the state-of-the-art, as it enables variable reaction time and easy catalyst switching/replacement. The process principle presented here is generic for liquid-phase reactions requiring long residence times, where conventional fixed-bed column reactors offer little flexibility. This is especially important when one switches between feed stocks or when the catalyst activity declines over time. Biodiesel manufacturing is a highly relevant example, because the reactor performance depends on the feedstock nature and composition. The concept is demonstrated in a scaleddown continuous laboratory reactor, keeping the same reaction time and comparable heat and mass transfer to the large-scale process by optimising the reactor dimensions, fluid velocity and catalyst pellet size. We then provide the design of the large-scale process, which consists of serpentine-type plug flow reactors assembled as vertical tubes filled with catalyst. The reactor productivity can increase significantly by reducing the catalyst pellet size. A switching system allows connecting/bypassing the tubes and easy catalyst replacement. The reactor can be employed in two-stage reaction technology, or in one-stage reaction combined with membrane separation. Production capacity can be scaled-up simply by adding parallel modules. The versatility and ease of application make this catalytic process concept suitable for low-cost mobile biodiesel production plants.

33

Abstract

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

1

Catalysis Science & Technology

Page 2 of 20 View Article Online

2

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Introduction Mankind’s short-term interest in biofuels and alternative energy sources might be affected by disorders in the spot price of crude oil, but in the long-term we will have to switch from fossil carbon sources to renewable ones. Biomass holds the key to the transition to a low-carbon economy, as it is widely available and renewable on a human timescale. Indeed, biorefineries will play an important role in any future scenario of sustainable energy management.1 Thus, developing new processes for converting biomass efficiently into fuels and chemicals should be a top priority.2 Biodiesel is a case in point. Today, two main types of compatible diesel fuel are commercially produced at large scale: fatty acid methyl esters (FAME) and hydroprocessed esters and fatty acids (HEFA). Following US National Biodiesel Board biodiesel and some experts3 the term of biodiesel refers primarily to FAME, made mostly by the transesterification of triacylglycerides (TAG) and/or by the esterification of fatty acids (FA). Biodiesel primarily feedstock can be vegetable oil crops, such as rapeseed (canola), sunflower, soy and palm (first generation), but preferably waste lipid material and non-edible vegetable oils, as jatropha, camelina and ricinus (second generation). Future resources with high potential are the algal biomass and the fermentation of carbohydrate wastes to lipids by special yeasts (third generation).4 HEFA fuel can be made by the hydrogenation of all lipid feedstock types. Note that HEFA should be integrated into a petro-refinery environment to ensure the availability of cheap hydrogen, fluid catalytic cracking (FCC) facility and suitable separation equipment. Since collecting waste oil in large amounts raises logistic problems and costs, the most economical feedstock for HEFA remains shipping vegetable oil from remote plantations (e.g. palm oil), raising questions about its overall sustainability. HEFA is designated as “green biodiesel” by some commercial producers. Beside bioethanol, biodiesel is one of the two top renewable fuels. Following the last report of Renewable Energy Network for the 21 Century (REN21) the annual world biodiesel production rose for more than twelve times from 2004 to 2014 to about 30 billion litres, which represents a share of 23% from the global biofuels. HEFA was about 3% at 4 billion litres. Europe produced 11.5 billion litres (40%), while USA 4.7 billion litres (16%). Rapid growth is expected in the next years from South-Asia, South-America and African countries.5 The sustainability of biofuels is a moot point. Among the assessment criteria the most important issues are the greenhouse gas (GHG) impact, food security, indirect land use change (ILUC), socio-economic development and energy independence. Regarding the second aspect, there is no clear evidence that the biofuels market threatens the food supply chain. For example, in Europe the crops for biofuels, mostly biodiesel, use only 1% from the agricultural land.5 Moreover, reducing GHG with respect to fossil fuels is considered by most analysts as the key factor for sustainability. The GHG assessment performed by national and international agencies follow at present well-to-wheel LCA procedures. The model usually includes emissions from feedstock cultivation (fuels and fertilizers), transportation, and fuel production process.5-7 Results typically depend on the local conditions. For biodiesel, the GHG reduction varies between 30–90% (excluding ILUC).8 The upper level regards waste and non-edible oils, the mid level efficient energy crops, while the low level refers to some crops issued from non-sustainable farming. The most ambitious sustainability targets are

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

1

DOI: 10.1039/C6CY00426A

Page 3 of 20

Catalysis Science & Technology View Article Online

DOI: 10.1039/C6CY00426A

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

applied in Europe following the Renewable Energy Directive (RED).6 The GHG threshold in 2015 for biofuels was set at 35% with respect to the reference value of 83.8 g CO2-eq/MJ, which should increase to 50% in 2017 and 60% in 2018. In USA the EPA-RFS2 regulation sets the GHG reduction target at 20% for conventional renewable fuels, and at 50% for advanced renewable fuels.9 These rules shall stimulate the progress in technology and the advent of new sustainable feedstock sources. Note that sustainable biodiesel requires less GHG from the manufacturing process too. Adopting the technology of heterogeneous catalysis, which offers substantial saving in equipment, energy, and materials, can greatly help achieving this goal. Another step forward would be extending this approach to small (mobile) and medium plants by a new reactor technology, as we propose here. Biodiesel production by homogeneous catalysis and batch processes dominates today’s market. It uses relatively simple technology, easy adaptable to different feedstocks, and suitable for small and medium-capacity plants. Among recent progress in technology the most potent is employing solid base catalysis.10 This can reduce substantially both capital and operation costs, delivering high-purity glycerol as a bonus. The French company AXENS developed the EsterfipTM process based on a zinc aluminate catalyst.11-13 The preferred feedstock is vegetable oil with low free fatty acid (FFA) and water content. This process was implemented in large-scale plants in France, Malaysia, Sweden and the US. Six plants were built up to 2012 with a total capacity topping 1.2 Mtpa. These plants use fixed-bed adiabatic tower reactors with the catalyst split into two sections, and provided with devices for ensuring homogeneous mixture and plug-flow profile.13 The design and operation of such reactors is constrained by the nature and the quality of the feedstock. Catalyst robustness is a key issue here, since the process uses large amounts of catalyst. A long time-on-stream is needed, preferably over six months. Furthermore, compensating the catalyst deactivation by raising the temperature is limited, due to the risk of glycerol decomposition. Thus, the challenge of the catalytic process is dealing with diverse industrial-grade feeds and converting them efficiently into a product respecting strict quality specifications, as defined by EN14212 and ASTM D6751. In 2006, our group published the first example of biodiesel synthesis from FFA using heterogeneous catalysis14, extending this later to various feeds and process options.15, 16 Here we present a solid catalyst process combined with an innovative reactor design that offers flexible reaction times and easy catalyst change upon deactivation. Importantly, this system is modular, and therefore suited for smaller scale and mobile production plants. This makes it compatible with the requirements of biorefinery concept, as it can valorise locally available resources with low investment and operation costs.

37

Solid catalyst technology

38 39 40 41 42 43 44 45

The manufacturing of biodiesel by heterogeneous catalysis has major economic benefits compared to the traditional homogeneous catalysis. As shown in Fig. 1, the main reason is the suppression of operations generating large amounts of waste water (except during feedstock pretreatment), namely the catalyst removal by acid/base neutralisations, washing of FAME and glycerol, as well as at methanol recovery by distillation of aqueous solutions. There is no salt waste, and catalyst consumption drops dramatically. In the homogeneous process the glycerol has maximum 85% purity after expensive separations, while the heterogeneous

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

3

Catalysis Science & Technology

Page 4 of 20 View Article Online

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

DOI: 10.1039/C6CY00426A

process delivers glycerol that is >98% pure, a valuable product. In addition, the continuous operation is intrinsically more efficient than batchwise.17 Heterogeneous catalysis may be applied both to low and to high FFA oils. In the last case a preliminary esterification with methanol is necessary, which can be performed by employing solid acid catalyst, as Amberlyst-type ion-exchange resins.3,17 The small amounts of water formed can be removed simply by adsorption. Higher FFA lipids, as animal fat and industrial greases, can be treated in a reactive distillation device.18 Benchmark calculations performed at Yellowdiesel B.V19 using a factorial method based on the purchasing cost of basic equipment20 indicate that for a plant of 100 ktpy the capital expenses (CAPEX) can drop by 40–60% since less equipment and more compact design. In addition the operating expenses (OPEX) can fall by 40-50% mainly by deleting energy intensive operations involving water, but also by a significant saving in the catalyst cost. Note that the higher temperature in the reaction stage, around 200 °C, should not penalise the energy consumption if adequate heat integration measures are taken.17, 20 These include feed preheating by reactor effluent and employing thermal fluids instead of steam. Similar economic benefits were reported by comparing homogeneous and heterogeneous biodiesel processes by simulation with Hysys.21

Figure 1 - The advantage of biodiesel manufacturing by heterogeneous base catalyst

Figure 2 shows a conceptual flowsheet inspired by the original EsterfipTM process.11-13 Methanol and vegetable oil pumped at higher pressure heated up and mixed enter the first adiabatic transesterification reactor (R-1). The preferred operation conditions are pressure between 40–70 bar, temperature from 200–220 °C, liquid hourly space velocity (LHSV) of 0.5–1 h–1, and a MeOH:oil weight ratio of 1:2. The catalyst is extrudates of 3 mm diameter. After pressure reduction, the outlet mixture is flashed for removing the excess methanol such to allow easy separation of glycerol by decantation, needed for shifting the chemical equilibrium. The ester phase enters a second reaction stage (R-2) similar with the first step, so methanol make-up and feed preheating are needed. In the first stage the conversion may rise up to 90-93%, while in the second stage up to 99.5%. The final FAME mixture is submitted to methanol removal by two-stage evaporation, followed by glycerol separation by decantation. The excess methanol is recycled after the distillation of glycerol and FAME streams. Note that in the above process the purity specifications of biodiesel, namely with respect of residual TAG, should be achieved almost completely after the reaction section. Thus, the

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

4

Page 5 of 20

Catalysis Science & Technology View Article Online

1 2 3 4 5 6 7 8 9 10 11 12 13 14

15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

DOI: 10.1039/C6CY00426A

reactor design should provide sufficient oversizing for dealing with feedstock variability, catalyst deactivation, and other disturbances. When using the zinc aluminate catalyst, the water content of feedstock should be limited to 500 ppm, and FFA acidity index lower than 10 mgKOH/g oil.11 A shortcut calculation brings an useful insight into the reactor design. Let us consider a production rate of 150 ktpa, 8000 hours annual operation, 90% yield, and 600 kg/m3 the density of oil/methanol mixture. It results a reactor volume of 74.4 m3. This value will be confirmed later in this paper by more accurate computation based on detailed kinetics.26 If we assume a reactor diameter of 2.4 m, the height of the cylindrical part hosting the catalyst should be at least 16.5 meter. The fictive fluid velocity, based on the empty cross section would be about 3.2 mm/s or 11.5 m/h. In addition, the plug flow profile requires adequate internals for mixing and redistribution, as well as for catalyst handling. Designing such a unit in practice is further complicated by the uncertainty regarding the kinetic behaviour of the feedstock and by catalyst robustness.

Figure 2 – Conceptual flowsheet for FAME manufacturing by heterogeneous catalysis

The EsterfipTM catalyst was discovered by Stern et al.22 in 1999 and improved in subsequent patents.23,24 The generic formula is ZnAl2O4.xZnOyAl2O3 in which x and y are numbers in the range 0 to 2. Recent developments describe a spinel-type catalyst robust and efficient because very little zinc leaching.25,26 However, the robustness in industrial conditions was not disclosed. The success of EsterfipTM led an intensive research in the field of solid base catalysts. Comprehensive reviews were published.27,28,29 Mixed metal oxide catalysts, namely hydrotalcites, showed good activity at higher temperature and reasonable reaction time.30-34 Note that dolomites, comparable as structure to hydrotalcites, could be used as abundant lowcost catalyst.10,35 Among latest developments one can cite the double-metal cyanide Fe-Zn catalyst developed by Indian Pune Laboratory36, which proved to be successful both for the esterification and transesterification of oils containing a significant amount of FFA. A process using this catalyst was implemented as ENSEL® process37 by the American company Benefuel. Despite abundant research, the literature study indicates that the synthesis of an efficient and robust solid catalyst for the transesterification of TAG remains a considerable challenge.

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

5

Catalysis Science & Technology

Page 6 of 20 View Article Online

6

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

Experimental: A miniplant for biodiesel manufacturing

Figure 3 – Schematic of a miniplant for continuous biodiesel manufacturing by heterogeneous catalysis

When dealing with solid catalysts for liquid-phase reactions, batch micro-reactors are used in most academic research for preliminary investigation. The catalyst is usually employed as powder. This method focuses on chemistry by eliminating mass and heat transfer resistances. The time evolution of concentrations in a batch reactor mimics the concentration profiles along the length of an ideal PFR. However, batch experiments are not sufficient for industrial application. The investigation should use an experimental device offering hydrodynamic similarity close to the industrial implementation. Accordingly, the pellet size should be adapted to the internal and external mass and heat transfer. Appendix 1 presents a simple method for designing a laboratory PFR reactor by employing similarity principles. It is important to note that the scale-down in terms of diameter and length are very different. Typically the diameter ratio could be over 100:1, while the length ratio only up to 10:1. With these elements in mind, we designed and built a mini-plant for manufacturing commercial grade biodiesel by heterogeneous catalysis (see schematic in Fig. 3). The chemical reactor is a spiral coil from SS 316 of ID 6 mm and OD 9 mm, with lengths of 2.4 and 3.6 m, filled with 0.8 mm catalyst. A 1 m silicagel pre-column is used for feed drying. The reactor is hosted in an electrical oven (Heraeus; Germany) with the internal size of 60×55×55 cm. The electronic regulated temperature can rise up to 250 °C. Oil and methanol are fed through metering membrane pumps (LEWA; Germany) with a maximum flow rate of 250 ml/h and 50 bars. The oil/methanol mixture is preheated and homogenised before reaction by passing it through a static mixer. A central element of the plant is the backpressure device, which has to ensure single liquid phase well below the saturation (bubble point). For this reason the pressure is set at minimum 5 bar higher than the methanol vapour pressure, namely in the range 35–50 bar, corresponding to reaction temperatures from 180–210 °C. The above set-up is capable for ensuring variable reaction time, in general LHSV 0.5–2 h–1, by connecting several reaction coils in series, or by adjusting the feeds of reactants. Flushing the reactor with methanol flow is used both for starting-up and shutting down the plant. The analytical equipment was Thermo Science GC apparatus with PVT/SSL backflush injector and FID detector equipped with Restek-Stabilwax column, which allowed accurate identification of methyl esters from C14 to C22. In view of GC analysis pure C14 was added in feed as 10 vol% with respect to oil. The above described miniplant was used for investigating potential industrial feedstocks. Pretreated feedstock as rapeseed, jatropha and frying fat were supplied graciously by the biodiesel producer Solarix B.V. Amsterdam, while sunflower, soya, and peanut were commercial vegetable oils. To prevent any influence of the impurities on the kinetic

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

1

DOI: 10.1039/C6CY00426A

Page 7 of 20

Catalysis Science & Technology View Article Online

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

20 21 22 23 24 25 26 27 28 29 30 31 32 33

DOI: 10.1039/C6CY00426A

behaviour, the feedstock was additionally cleaned by treatment with phosphoric acid 40 w% solution, neutralised, decanted, washed with distilled water and dried under vacuum. The experiments reported here employed an MgAl hydrotalcite catalyst supplied by Eurosupport B.V. Nederland as cylindrical pellets of 1.5×3 mm. The catalyst was crushed, sieved and flushed with nitrogen for getting a uniform size of 0.8 mm free of fines. The hydrotalcite catalysts have as remarkable feature a layered-type arrangement, which is advantageous for the internal diffusion of bulky TAG molecules, as well as for the extraction of the reaction products. The catalyst structure was characterised by employing a Thermo Fischer Scientific mercury porosimeter, and is reported elsewhere38. Key results are: bulk density 1.04 g/cm3, void volume 680 mm3/g, accessible porosity 58%, internal area 90 m2/g, average pore diameter 25 nm, median pore diameter 63 nm, and maximum pore diameter 275 nm. For comparison, a catalyst used in the Esterfip process25 had a specific surface area of 160 m2/g and mesopores ranging between 9 and 100 nm. The availability of very large pores favours the internal diffusion of the bulky TAG molecules, and support complex surface reactions for converting TAG to FAME. In addition, the talcite catalysts can be produced from low-cost clay materials. The regeneration can take place by simple calcination at temperatures between 500 and 600 °C. The waste catalyst is a nontoxic material that can be used as construction filler.

Figure 4 Continuous transesterification run in a laboratory fixed-bed reactor

Figure 4 shows the TAG conversion over 24 h TOS in a reactor of 3600 mm length filled with 103 ml catalyst at 190 °C and 37 bar using rapeseed oil as feed. The ratio methanol/oil is 0.55 and the combined feed of 175 ml/h. It results LHSV of 1.72 h-1, or a residence time of 35 min, with a fictive velocity of 1.7 mm/s. The conversion is practically constant around 68% indicating unchanging catalyst activity. After methanol evaporation and glycerol decantation, the transesterification can be continued in a second stage, where the conversion can be pushed to > 90%, and then in a third run to over > 98%. By decreasing the LHSV to about 1 h-1 the conversion in the first run can rise to about 90%, requiring only two transesterification stages. The catalyst activity remains nearly constant over several days with runs of 10 hours, the reactor being flushed at shut-down with methanol minimum two times the residence time.

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

7

Catalysis Science & Technology

Page 8 of 20 View Article Online

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

1 2 3 4

However, at longer operation time the activity declines slowly. The catalyst can be regenerated by simple calcination in an electrical oven at 550 °C, recovering at least at 90% of its activity. This behaviour in terms of activity and stability is close to the results reported recently with hydrotalcite catalyst in fixed-bed reactor over 55 h TOS.39

5

Results and Discussion

6

The influence of feedstock on the reactor performance

7 8 9 10 11 12 13 14 15 16 17 18 19 20

The feedstock for biodiesel shows a large compositional variation. Table 1 illustrates typical oil composition expressed as the amounts of saturated and non-saturated fatty acids. Data for jatropha correspond to South-American crops.40 As indicated, saturated C16 and C18 fatty acids dominate in palm kernel oil, tallow and frying fats. Rapeseed is richer in monounsaturated (MUS) oleic acid. Peanut and soya oils have large amounts of poly-unsaturated (PUS) as C18:1 and C18:2, but soya contains also a substantial amount of C18:3. Jatropha oil is richer in C18:2, together with substantial amounts of C16:0 and C18:1. Table 1 displays also the ratio R of unsaturated/saturated fatty acids, which varies from 0.16 for palm kernel oil to 13.3 for rapeseed oil. The spatial conformation of triglyceride molecules is very diverse, resulting from the shape of fatty acids. This can evolve from straight chains to more or less squeezed/bend-shape molecules. Accordingly, the diffusion properties of triglyceride molecules through porous solid catalysts should be very different too. Fatty acid type composition (% weight)

Feedstock type Palm kernel oil Frying fat Pork tallow Jatropha Peanut oil Soya oil Rapeseed 21 22 23

C16:0 80 50 10 22 10 10 4

C18:0 6 7 41 6 4 4 3

C20+ 0 0 0 1 5 0 0

C18:1 12 35 42 22 40 25 65

Table 1 – Typical composition of triglyceride oil feedstock for biodiesel

C18:2 2 8 3 48 41 51 20

C18:3 0 0 4 1 0 10 8

R 0.16 0.75 0.96 2.45 4.26 6.14 13.29

Catalysis Science & Technology Accepted Manuscript

DOI: 10.1039/C6CY00426A

8

Page 9 of 20

Catalysis Science & Technology View Article Online

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

DOI: 10.1039/C6CY00426A

Figure 5 – Comparative experiments in batch reactor

As demonstrated in an early paper38, when using a hydrotalcite catalyst there is a relation between the chemical composition of the feedstock, namely the ratio saturated to nonsaturated fatty acids, and the kinetics of the transesterification reaction, with important consequences on the chemical reactor design. Figure 5 presents curves time-conversion with four typical feedstock, frying fat, soya, peanut and rapeseed oils. The ordinate represents the total esters concentration in GC units. The experiments were conducted in batch mode in a high pressure stirred autoclave of 150 ml, at 190 °C and 40 bar. A typical experiment involves 100 ml oil, 25 ml methanol, 10 ml tetradecane as internal standard and 14 g catalyst. It can be observed that the kinetic behavior of oils is notably different. The first 45 minutes are in the heating period. The slope of the curves shows that the frying fat exhibits the fastest reaction rate, followed by rapeseed oil, soybean and peanut oils, the last one with a somewhat longer sluggish period. Considering only the linear range of the isothermal data gives a relative reaction rate of frying fat/rapeseed/soy/peanut as 1.6/1.26/1.17/1.0. These results suggest that the chemical nature of the feedstock may have an influence on the kinetics of the transesterification reaction. Faster reaction rates should be expected by higher content in saturated fatty acids, while the opposite should happen by unsaturated oils. Indeed, the behaviour of frying fat (R=0.75) seems to confirm this assumption. However, rapeseed oil (R=13) reacts somewhat faster than soy oil (R=6), while peanut oil (R=4) is the slowest. Table 1 shows that rapeseed oil has a higher content in C18:1 oleic TAG, soy oil has a higher content of C18:2 linoleic TAG, while peanut oil contains a notable amount of C20+ TAG. These results suggest that there are differences in kinetics among the unsaturated TAG too. The above kinetic behaviour may be explained by steric hindering effects on pore diffusion, as well as by interactions between the unsaturated bonds and the catalyst surface. The effective diffusion property of the TAG depends strongly on the chain’s shape. Thus, saturated TAG exhibiting “straight” chains, as palmitic C16:0 and stearic C18:0, should give easier diffusion inside the planar multi-layer structure. The cis-C18:1 oleic acid chains (Ω-9) show a ‘‘bent’’ shape and slower diffusion. The C18:2 linoleic (Ω -6) and C18:3 α-linolenic chains (Ω -3) show a “hook” shape that is even more likely to give steric hindering. From surface reaction viewpoint, the poly unsaturated chains may be slow-down by hydrogencatalyst bonds. Very long C20+ chains could give even slower diffusion by bending the molecule to access the reaction site.

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

9

Catalysis Science & Technology

Page 10 of 20 View Article Online

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

23 24 25 26 27 28 29 30

DOI: 10.1039/C6CY00426A

Figure 6 – Kinetics of transesterification for individual triglycerides

A more accurate description of the kinetic effects was obtained by examining the behaviour of the individual fatty acids. Fig. 6 displays the curves concentration vs. time by soya oil transesterification in a batch reactor as described before, by monitoring the formation rate of for groups of esters, namely palmitic plus stearic, oleic, linoleic and linolenic. The ratio of methanol/oil was 0.5 w/w or about 15:1 mol/mol. Since large excess of methanol, first order kinetics can be safely assumed and the following relation applies:

kτ = − ln (1 − X / Xeq )

(1)

where k is the first-order reaction constant, τ the reaction time, X the actual conversion and Xeq the equilibrium conversion. Theoretically k=k1+k-1 where k1 and k-1 are the kinetic constants for forward and backward reactions. Since the equilibrium constant is large, of about 50 (see Table 3), one gets k≈k1. Figure 7 presents the results of a linear regression model of the isothermal data from Fig. 6, namely at times longer than 60 minutes. Table 2 gives the values of the kinetic constants and the mean errors of fitting. Table 2 – First order reaction constant transesterification of individual triglycerides C16:0 & C18:0 C18:1 C18:2 C18:3 k, min-1 0.045±0.0023 0.0305±0.0012 0.0268±0.0015 0.0239±0.0014 k ratio 1 1.47 1.68 1.88 The results show that the catalytic reaction rate of different TAG depends on the type of fatty acid. Saturated chains as palmitic and stearic TAG exhibit the highest rate. The oleic TAG reacts about 47% slower, showing that the double bond introduces a significant molecular hindering effect. It follows linoleic and linolenic TAG, with additional delays. The question arises if these differences in reaction kinetics may have practical consequences for the design and operation of biodiesel processes, namely when using feedstock with different TAG profile.

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

10

Page 11 of 20

Catalysis Science & Technology View Article Online

11

DOI: 10.1039/C6CY00426A

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Figure 7 – Linear regression of kinetic data for individual triglycerides

By using the composition given in Table 1 and kinetic values from Table 2, average first order constants have been calculated, and then the reaction time for given conversion was reconstructed. Fig. 8 presents curves for industrial relevant feedstock, as rapeseed, soybean, jatropha and frying fat. Two reference curves were also traced for palmitic (C16:0) and linoleic (C18:2) TAG. The results show again a significant difference in kinetics. Faster runs are obtained with essential saturated oils, as kern palm oil and coconut oil, while unsaturated oils, as rapeseed, soybeans and jatropha, show a slower kinetics, dominated by the presence of oleic and linolenic chains. Frying or tallow fats, including in comparable proportion saturated and unsaturated TAG, have an intermediate kinetics. The differences between the kinetic behaviour widen with increasing conversion. Rapeseed and soya oil shows similar kinetics, which can be explained by the compensating effect of more saturated TAG in the last. Note that the above computations take into account only the key fatty acids, from C16 to C18. The presence of very large TAG molecules, ignored in this study, and of impurities, which may be trapped in the catalyst structure, might add supplementary slow-down effects on the kinetics of reaction. The results reveal two important technological aspects. The first one regards the inter-stage separation of glycerol. For applying two-stage technology the TAG conversion in the first step should be around 90%. Fig. 8 indicates that the necessary time is 60 minutes for frying fat, but 75 minutes for rapeseed oil. After one hour run the conversion for rapeseed oil would be only 84%, insufficient for getting 98.5% in a second stage. A substantial increase of the reaction time, or a third stage would be necessary. If the reactor has only limited oversizing capacity this behaviour becomes problematic. The same troubles occur when the catalyst activity declines. The second aspect is related to productivity. Since very high conversion is needed for respecting quality specifications, this is easier achieved with saturated TAG. Again, when the conversion in the first stage is insufficient, a third or a fourth reaction stage is needed, and as consequence the productivity drops drastically.

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

1

Catalysis Science & Technology

Page 12 of 20 View Article Online

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

1 2 3 4 5

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

DOI: 10.1039/C6CY00426A

From the above investigation, it is clear that the reactor design for biodiesel manufacturing by heterogeneous catalysis should provide sufficient flexibility both in ensuring adaptable reaction time to the feedstock type, and for easier catalyst replacement when activity falls, by leaching or by contamination. This is the object of the next section.

Figure 8 - Relation between the chemical nature of feedstock and the reaction kinetic

An innovative reaction system for catalytic biodiesel manufacturing Figure 9 presents the innovative solution proposed by us in a patent application.41 The variable reaction time device consists of a serpentine-type plug flow reactor, assembled as vertical tubular segments filled with solid catalyst. A switching valve system (not shown) is employed for connecting or bypassing the reaction tubes, as well for easy catalyst replacement. The set-up makes use of static mixers. Heating & cooling elements are provided for mixture conditioning before and after reaction. Saving energy is obtained by countercurrent coupling as feed-effluent heat exchangers (FEHE). A liquid thermal agent (Dowtherm, hot oil, etc.) is used as heat carrier. The transesterification takes place in two stages with intermediate glycerol removal. The glycerol separator can be a gravity coalescence separator making of hydrophobic/hydrophilic materials to accelerate the formation of larger drops42, or a compact centrifugal device. The final conversion should be higher than 98.5 %. The operation conditions for hydrotalcite catalyst are temperature of 180–220 ºC, pressure 35–70 bars, and LHSV of 0.5–2 m3/m3 catalyst/h. The construction allows adjusting the residence time to the feedstock type and to the catalyst activity by varying the number of the active tubes. Moreover, the scale-up of plant capacity can be done easily by assembling parallel “reaction boxes”. Note that the above presented set-up is generic for liquid-phase reactions that need larger volume, longer residence time, and making use of solid catalyst. Double-pipe heat exchangers can ensure an effective heat transfer in the zones where this is needed, for re-boosting the chemical reaction.

Catalysis Science & Technology Accepted Manuscript

12

Page 13 of 20

Catalysis Science & Technology View Article Online

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 9 - Compact reaction/separation system using tubular segmented reactor for biodiesel

manufacturing

Another innovative solution makes use of only one-step transesterification in variable reaction time device coupled with membranes separation, which ensures recycling both triglycerides and methanol. Commercially available ultrafiltration and nanofiltration membranes can easily separate between small and large molecules under these conditions.43 In biodiesel manufacturing there are significant differences between the molecules both in size and shape, as well as in the hydrophobic / hydrophilic character. As order of magnitude, the kinetic diameter is about 0.4 nm for methanol and 0.6 nm for glycerol. The kinetic diameter of fatty acid alkyl esters depends on the length and character of the hydrocarbon chain, and is generally in the range of 0.8 to 1.5 nm. The size of triglyceride molecules is larger by almost an order of magnitude. However, there is other physical phenomenon that can be exploited for separation, as shown in Fig. 10. In immiscibility conditions an emulsion can form, in which the TAG molecules surrounded by methanol are segregated in large particles. This possibility was demonstrated experimentally by the transesterification of canola with methanol; oil droplets of 44 µm mean diameter were formed with size distribution from 12 µm to 400 µm.44

Glycerol TAG

MeOH

22 23 24 25 26 27 28 29 30 31

FAME

Figure 10 - Separation of triglycerides by membrane microfiltration

Fig. 11 shows the conceptual flowsheet. After transesterification and cooling, the outlet mixture well dispersed (pumping or static mixing device) is submitted to the first separation by microfiltration. TAG molecules remain in retentate, while FAME, glycerol and methanol pass in permeate. TAG are recycled to the catalytic reactor, a drain being provided to prevent accumulation of unreacted oil and impurities. For this separation commercial ceramic or carbon membranes are suitable with pores in the range 0.4 to 1.4 µm44. After remixing,

Catalysis Science & Technology Accepted Manuscript

DOI: 10.1039/C6CY00426A

13

Catalysis Science & Technology

Page 14 of 20 View Article Online

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

permeate is submitted to a second membrane separation, this time by ultrafiltration. FAME and glycerol are immiscible, but glycerol and methanol are miscible. In addition, the diffusion of glycerol through membrane will be hindered by higher viscosity. As result, methanol goes in large majority in permeate and is recycled to the reaction, while FAME/glycerol mixture remains in the retentate and is sent to separation by decantation or centrifugation. This method may be feasible with ceramic membranes with pores of 0.2 µm, by analogy with experiments by separating biodiesel, glycerol and ethanol.45 As support for the viability of the above approach, we may associate recent researches regarding membrane reactors for biodiesel working with homogeneous catalysis.46-48 More information on process intensification aspects at biodiesel manufacturing can be found in a recent monograph.49 Putting in a nutshell, the advantages of this method are: a. More flexibility in the reactor operation. The TAG conversion could be in a wide range between 70 to 90%, while high recovery of TAG and methanol in recycles is not required. This time the quality of biodiesel is decided in the separation section, not in the reaction steps, as previously. b. Important energy saving can be obtained by recycling most of the methanol in liquid phase and not by distillation. Only a limited amount of methanol carried with FAME and glycerol need to be recovered by distillation. Conversely, larger methanol/oil ratio could be employed for boosting the transesterification by internal recycling. c. The problem of soap formation, with very negative effect on the operation costs50 can find an efficient solution. d. Better specifications of the end-product are obtained. e. The pressure differences for driving the membrane process are available from the initial stream pressure. f. The design and operation parameters can be optimised on a wide scale as function of the reaction parameters and membrane characteristics. cooler V-1 Heater

Membr-1

STM V-2

FEHE

FAME+ GLY+ MeOH

Methanol recycle separation

Crude BD FAME + (MeOH)

Membr-2 Oil P-1 P-2

29 30 31 32 33 34 35 36 37 38

Mono-Di-Triglyceride recycle

FAME + GLY + (MeOH)

GLY + (MeOH) Methanol recycle reaction

Methanol

Figure 11 – Single-stage reaction using tubular segmented reactor and membrane separation for

biodiesel manufacturing

As an example, we present a comparative design of a chemical reaction system with intermediate glycerol separation for the Esterfip process employing a ZnAl2O4 catalyst. Table 3 shows kinetic data adapted from Allain et al.26 considering an overall efficiency of 0.25 that multiplies the pre-exponential factors of the original rate equations. This is necessary since the industrial reactors employed extrudes of 3.0 mm while the lab reactor was filled with grains of 0.4 mm. Glycerol is completely removed after the first reaction step and no

Catalysis Science & Technology Accepted Manuscript

DOI: 10.1039/C6CY00426A

14

Page 15 of 20

Catalysis Science & Technology View Article Online

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

26 27 28 29 30 31 32 33 34 35 36 37

DOI: 10.1039/C6CY00426A

supplementary MeOH is added in the second stage. The final product is free of methanol and glycerol. A Matlab® simulation using a pseudo-homogeneous concentration-based model was constructed. The key assumptions are: 1) plug-flow profile, no maldistribution of the reaction mixture and no radial dispersion since suitable hydraulic design and upward flow13; 2) negligible external resistance to diffusion since fast mass transfer compared with internal diffusion and reaction by long residence time; 3) negligible temperature gradients since small heat of reaction; 4) internal diffusion of TAG and FAME molecules is process rate controlling via the overall efficiency mentioned earlier. These assumptions are supported by a comprehensive 2D simulation model of an industrial reactor.51 The key role of the internal diffusion on kinetics is emphasized by the present study too. Thus this model allows getting useful insights into the catalytic reactor technology, as the relation residence time, catalyst activity and final product specifications. Following EN14214 minimum specifications are ester content 96.5%, triglycerides TAG 0.2%, diglycerides DAG 0.2%, monoglycerides MAG 0.7%. The reactors are identical with diameter and length of 2.4 and 16 m respectively, with a volume of 72.35 m3 each. The feed consists of 300 kmol/h methanol and 25 kmol/h trioleine, which gives a combined feed of 31745 kg/h, at a molar ratio of 12 or 0.434 weight ratio. The temperature is set constant at 190 °C. Considering a mean density of 600 kg/m3 gives a LHSV of 0.73 m3/m3 catalyst/h. The mean fluid linear velocity in the reactor is around 3.25 mm/s, which should ensure a plug flow profile and an acceptable mass transfer rate. Table 3 – Kinetic data for sizing the catalytic reactor (adapted from Allain et al.16) Reaction k0 m6kg-cat/kmol/s Ea kJ/kmol Keq 1 TAG+MeOH ↔ DAG+FAME 3.15E+00 64600 51.2 2 DAG+MeOH ↔ MAG+FAME 2.22E-03 31800 53.1 3 MAG+MeOH ↔ GLY+FAME 3.18E-05 17000 12.2 Figure 12 illustrates the dependency of reactor conversion as function of catalyst activity, as well as the evolution of product specifications. In the base-case, the final product rate is 22260 kg/h and accordingly the productivity over the two reactors is 153.85 kg/m3/h or 3692 kg/m3/day. The conversion in the first reactor should be over 88% to ensure an in-specs final product. If it drops below 80%, then the product is off-specs. Since the reactor volume is fixed, restoring the specifications can be done only by rising the temperature, for example from 190–205 °C when the activity falls to 0.6, or by reducing proportionally the throughput. Manipulating the reaction temperature is also technologically constraint to 15–20 °C and implies more energy consumption. Changing exhausted catalyst needs plant shut-down and throwing away the whole load.

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

15

Catalysis Science & Technology

Page 16 of 20 View Article Online

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

DOI: 10.1039/C6CY00426A

Figure 12- Reactor conversion and specifications plotted vs. catalyst activity

But when a serpentine-type reactor is employed, additional tubes can be connected to compensate the fall in catalyst activity. By monitoring the activity at selected locations it is possible to replace only the catalyst that really needs it. The scale-up of the production rate can be simply achieved by a modular design. Moreover, when using tubes smaller size catalyst can be employed, with the effect of increasing the reaction rate. For example the pellet size can be reduced from 3 mm to 1 mm, which can double the reaction rate, as predicted by efficiency calculation using pseudo-first reaction and Thiele modulus. Thus productivity of 300 kg/m3/h or 7200 kg/m3/day can be achieved. The individual reactor volume for the same production can be reduced from 72.35 to about 40 m3. This volume can be hosted in a reaction box comprising 8 modules of 5 m3, each consisting of eight tubes of 0.5 m diameter and 3.2 m length. It can be seen that this reactor design has much more favorable conditions for operation, as flexibility in feed and catalyst replacement. The compact reaction/separation system described above is particularly interesting for designing mobile biodiesel plants. The hardware is split in sections, as reaction, glycerol separation, methanol recovery, energy recovery, hot oil utility and catalyst maintenance. The production parts can be mounted on containers movable on trucks at the manufacturing locations. The storage tanks for raw materials, intermediates and products, implying large volumes and substantial investment, should remain on field. Deactivated catalyst can be recharged locally with ready-to-use tubes. This solution is convenient from the viewpoint of seasonal harvesting too. As example a device comprising two-stage reaction boxes - each one built as four serpentines with 15 tubes of 0.25 m ID and 2.5 m length and loaded with 14.7 m3 high productive catalyst of 300 kg/m3/h - could deliver about 100 tpd biodiesel.

28

Conclusions

29 30 31 32 33 34

The manufacturing of biodiesel by solid base catalyst emerged as advanced sustainable technology. Beside considerable reduction in capital and operation costs, pure glycerol is got as valuable by-product. The classical chemical reactor for such process is a tall fixed-bed large volume tower, whose design and operation is constraint by the influence of the feedstock on the reaction kinetics, as well as by the catalyst deactivation.

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

16

Page 17 of 20

Catalysis Science & Technology View Article Online

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

DOI: 10.1039/C6CY00426A

An experimental study performed in batch reactor and in continuous micro-plant with several types of oils - rapeseed, jatropha, frying fat, soya, peanut - demonstrates that there are kinetic effects related to the feedstock nature. The catalyst was of hydrotalcite-type, which presents a layered structure favouring an easier diffusion of large molecules. Kinetic constants of individual fatty acid triglycerides were evaluated for saturated (palmitic and stearic) and unsaturated (oleic, linoleic and linolenic) TAG. Saturated triglycerides give much faster reaction rate than unsaturated counterparts. The explanation lies in the fact that the internal diffusion is slowed down by both steric and chemical bonding effects related to the chemical character of TAG molecules comprised in the feedstock. It is worthy to note that recent papers published by French researchers using the Esterfip-type catalyst arrived at the conclusion that the internal diffusion of triglycerides is an important factor limiting the global conversion.26,51 A continuous micro-plant was built in view of an industrial process development. A laboratory reactor has been designed such to respect key similarity criteria with the industrial reactor. The scale-down is very different in terms of diameter and length. The lab reactor should be sufficiently long in order to ensure the same residence time, while keeping a suitable fluid velocity that preserves similar mass and heat transfer particle/fluid. Since the fluid velocity drops significantly in the lab device, the size reduction of pellets may be necessary too. The operation of the micro-plant with various raw materials demonstrated the ability of the hydrotalcite catalyst to handle various raw materials and produce quality biodiesel in two-stage run with a LHSV of about 1 h-1. The experience acquired inspired the design of an innovative set-up employing solid catalyst that solves two shortcomings of tower-type reactors: variable reaction time and easy catalyst change. The device consists of a serpentine-type plug flow reactor, assembled as vertical tubular segments filled with catalyst. A switching valve system is used for connecting or bypassing the reaction tubes, and for catalyst replacement. The residence time can be adapted to feedstock type and catalyst activity by varying the number of the active tubes. Moreover, the scale-up of capacity can be done easily by assembling parallel “reaction boxes”. Another valuable advantage is the significant increase in the reaction rate that can be obtained by using smaller catalyst pellets and better mass transfer conditions. Further compactness could be obtained in just one reaction stage by using a membrane device for separation and recycling unreacted triglycerides and methanol. The above solutions can be particularly interesting for building-up mobile biodiesel plants. The hardware module comprises reaction, glycerol separation, methanol recovery, and hot oil utility part, all movable on containers to the production location. This concept is suitable for biorefineries that above all should valorise local resources.

Acknowledgement This work is part of the Research Priority Area Sustainable Chemistry of the UvA, http://suschem.uva.nl

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

17

Catalysis Science & Technology

Page 18 of 20 View Article Online

DOI: 10.1039/C6CY00426A

18

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

6 7 8

Appendix: Design of a laboratory catalytic PFR device The similarity condition regarding conversion and yield is the equality of liquid hourly space velocities, or of the residence times, in the industrial and experimental reactors. This is LHSV1=LHSV2, or: Qv1

(π / 4) D12 L1

=

Qv 2

(π / 4) D22 L2

(A-1)

where Qv, D and L are the volumetric flow, reactor diameter and length. It follows that the scale-down of volumetric flow rates should respect the rule: 2

9 10 11

Qv 2  D2   L2  =    Qv1  D1   L1 

(A-2)

Accordingly, the scale-down of fluid velocity is: w2 L2 = w1 L1

(A-3)

12 13 14 15

Typically D1/D2 is over 100, while L1/L2 is below 10, and as result the fluid velocity could drop significantly in the lab reactor leading to low mass and heat transfer coefficients catalystfluid. For compensating this effect, reducing the pellet size is necessary. Let consider the mass transfer. This can be described by a Froessling type equation as Sh = 2 + Reαp Sc β , where the

16

Sherwood number is Sh = km d p / DA , Reynolds particle number Re p = wd p / ν , and Schmidt

17 18 19 20

number Sc = ν / D A . The notations signify km the mass transfer coefficient (m/s), dp the pellet equivalent diameter, DA the external diffusion coefficient of A species, and ν the fluid kinematic viscosity. It results that the following relation can be used for estimating the effect of fluid velocity and particle diameter on the rate of mass transfer:

21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

(

km = Cwα d αp −1 ≅ C w / d p

)

0 .5

(A-4)

since usually α≈ 0.5. Similar equation can be obtained for the partial heat transfer coefficient catalyst-fluid. The relations (2) to (4) can be used as a shortcut method for designing a laboratory fixed-bed tubular reactor that approaches a large-scale fixed bed reactor. As numerical example we take the above industrial catalytic reactor. The production rate is 150000 tpy for 8000 hours of continuous operation. The LHSV is 0.7 h-1, while the feed ratio methanol/oil is 0.5. The yield of TAG to FAME is 90%. The industrial catalyst consists of pellets of 3 mm equivalent diameter. For the laboratory reactor a tube of 6 mm ID is selected. The ratio of diameters is D1/D2=220/6=400. Assuming a fictive fluid velocity of 1 mm/s the tube length of 5.14 m would ensure the same LHSV as in the industrial reactor. The length ratio becomes L1/L2=16.5/5.14=3.2. The similarity of mass and heat conditions can be fulfilled if the pellet size is reduced to 0.94 mm. The combined oil and methanol flow for the laboratory reactor would be about 100 ml/h.

Catalysis Science & Technology Accepted Manuscript

1 2 3 4 5

Page 19 of 20

Catalysis Science & Technology View Article Online

19

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

Z. Strassberger, S. Tanase and G. Rothenberg, RSC Adv., 2014, 4, 25310-25318. R. Beerthuis, G. Rothenberg and N.R Shiju, Green Chem., 2015,17, 1341-1361. J. van Gerpen and G. Knothe, Basics of transesterification reaction, in The biodiesel handbook, 2005, AOCS Press. C. Perego and M. Ricci, Catal. Sci. Technol., 2012, 2, 1776–1786. REN21, Renewable 2015 Global status report. Available at http://www.ren21.net. Last accessed 5 April 2016. European Biofuels Technology Platform (EBTP), Factsheets. Available at http://www.biofuelstp.eu. Last accessed 5 April 2015. A. Azapagic and H. Stichnote, Life cycle sustainability assessment of biofuels, in Handbook of biofuels production, 2011, Woodhead Publ., UK. Wikipedia, Environmental impact of biodiesel. Available at https://en.wikipedia.org. Last accessed 4 April 2016. US EPA, Final renewable fuels standards for 2014 to 2017. Available at https://www.epa.gov. Last accessed 5 April 2016. A.F. Lee. and K. Wilson, , Catalysis Today, 2014, 242, 3-18. L. Bournay, G.Hillion, P. Boucot, J. Chodorge, C. Bronner and A. Forestiere., U.S. Patent 2005, No 6,878,837. L. Bourney, D. Casanave, B. Delfort, G. Hillion and J.A. Chodorge, Catal. Today, 2005, 106, 109-115. M. Bloch, L. Bournay, D. Casanave, J.A. Chodorge, V. Coupard, G. Hillion and D. Lorne, Oil & Gas Science and Technology – Rev. IFP, 2008, 63, 405-417. A.A. Kiss, A.C. Dimian and G. Rothenberg, Adv. Synth. Catal., 2006, 348, 75-81. A.A. Kiss, F. Omota, A.C. Dimian and G. Rothenberg, Top. Catal., 2006, 40, 141-150 A.A. Kiss, A.C. Dimian and G. Rothenberg, Energy & Fuels, 2008, 22, 598-604. A.C. Dimian and S. Bildea, Biodiesel Manufacturing, in Computer-Aided Design Case Studies, 2008, Wiley. F. Omota, A.C. Dimian and A. Bliek, Chem. Eng. Sci., 2003, 58, 3175-3185. A.C. Dimian, G. Rothenberg and R. Schut, Internal report Yellowdiesel B.V.-The Netherlands, 2010. A.C. Dimian, Integrated Design and Simulation of Chemical Processes, 2003, Elsevier, CACE-13. A. West, D. Posarac and N. Ellis, Bioresource Techn. 2008, 99, 6587–6601. Stern R., Hillion G. and Rouxel J.J., US Patent, 1999, No 5,908,946. G. Hillion, S. Leporq, D . LePennec and B. Delfort, US Patent, 2004, No 0234448; see also US Patent 2005/0113588A1. D. Bazer-Bachi, V. Coupard. S. Maury., V. Pugnet, I. Clemencon and A.A. Quoineaud, US patent, 2011, No 0092730 A1. V. Pugnet, S. Maury, V. Coupard, A. Dandeu, A. Quoineaud, J.L. Bonneau and D. Tichit, Applied Catalysis, 2010, 374, 71–78. F. Allain, J.F. Portha, E. Girot, L. Falk, A. Dandeu and V. Coupard, Chemical Engineering Journal 2016, 283, 6833–845. M. Di Serio, R. Tesser, L. Pengmei, and E. Santacesaria, Energy & Fuels, 2008, 22, 207–217. Y.C. Sharma., B. Singh and J. Korstad, Fuel, 2011, 90, 1309–1324. A.F. Lee, J.A. Bennett, J.C Manayil and K. Wilson, Chem. Soc. Rev., 2014,43, 7887.

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

1

DOI: 10.1039/C6CY00426A

Catalysis Science & Technology

Page 20 of 20 View Article Online

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

DOI: 10.1039/C6CY00426A

30 D.G. Cantrell, L.J. Gillie, A.F. Lee and K. Wilson, Applied Catalysis A: General, 2005, 287, 183. 31 N. Barakos, S. Pasias and N. Papayannakos, Bioresource Techn., 2008, 99, 5037–42 32 G.S. Macala, A.W. Robertson, C.L. Johnson, Z.B. Day, R.S. Lewis, M. White, A.V. Iretskii and P.C. Ford, Catal. Letters 2008, 122, 205–209. 33 C.M. Silva, N.F.P. Ribeiro, M. Souza and D. Aranda, Fuel Proc. Techn., 91, 205-210 34 M. Di Serio, S. Mallardo, G. Carotenuto, R. Tesser and E. Santacesaria, Catalysis Today, 2012, 54–58. 35 K. Wilson, C. Hardacre, A.F. Lee, J. Montero, L. Shellard, Green Chem., 2008, 10, 654-659. 36 P.S. Sreeprasanth, Srivastava R., Srinivas D. and Rattnassamy P., Appl. Catal. A: General, 2006, 314, 148–159. 37 D. Srinivas and P.Rattnassamy, US Patent 2012 No 8124801. 38 A.C. Dimian, Z. Srokol, M. Mitelmeijer and G. Rothenberg, Topics in Catalysis, 2010, 53, 1197-1201. 39 P. Kutálek, L. Čapek, L., Smoláková, D., Kubička, Fuel Processing Technology, 2014, 122, 176-181. 40 N. Tapanes, D.A.G Aranda, J. de Mosquito Carneiro, O. Ceva Antunes, Fuel, 2008, 87, 2286-2295. 41 A.C. Dimian, G. Rothenberg and R. Schut, European Patent, 2012, EP 2457648 A1 / WO 2012072574 A1. 42 V. Jordan, B. Gutsche, Chemosphere, 2001, 43, 99-105. 43 A.V. Gaikwad, V. Boffa, J.E. ten Elshof and G. Rothenberg, Angew. Chem. Int. Ed., 2008, 47, 5407-5410. 44 P. Cao, Y. André, M.A., Dubé, A.Y. Tremblay and M. Katie, Ind. Eng. Chem. Research, 2007, 46, 52-58. 45 M. C. Sérgi Gomes, N.C. Pereira and S.T. Davantel de Barros, J. Membrane Sci, 2010, 352, 271–76. 46 P. Cao, M.A. Dubé, Y.T. André, Biomass and Bioenergy, 2008, 32, 1028-36. 47 P. Cao, M.A. Dubé and A.Y. Tremblay, Fuel, 2008, 87, 825-33. 48 M.A. Dubé, A.Y Tremblay and J. Liu, Bioresource Technology, 2007, 98, 639-647. 49 A.A. Kiss, Process intensification technologies for biodiesel production, Springer Briefs in Applied Sciences and Technology, 2014. 50 I.M. Atadashi, M.K Aroua. and A. Abdul Aziz., Renewable Energy, 2011, 36, 437-443. 51 J.F. Portha, F. Allain, V. Coupard, A. Dandeu, E. Girot, E. Schaer, and L. Falk, Chemical Engineering Journal, 2012, 207-208, 285-298.

Catalysis Science & Technology Accepted Manuscript

Published on 27 April 2016. Downloaded by Universiteit van Amsterdam on 27/04/2016 11:02:18.

20